如何在图表中添加过滤器

Question

代码如下

from io import StringIO
text = '''Product,Count
Pen,10
Pencil,15
Book, 10'''
df = pd.read_csv(StringIO(text))
df.plot(x="Product", y="Count", kind="bar")

• 如何在图表本身中添加过滤器，用户必须有权选择哪些product必须显示在图表中，count并且假设count > 11只有 Pencil 必须出现。

• 有没有其他方法可以做到这一点？

• 如果一列是日期列，我们是否也可以使用日期列进行过滤

Answer 1

matplotlib.widgets

正如评论中所建议的那样，一种方法是使用matplotlib.widgets，您可以在此处阅读有关它们的更多信息，但对于实际实现，我发现它们的滑块和检查按钮示例最有用。使用您的最小示例，我能想到的最简单的改编（看起来不错）看起来像这样：

import pandas as pd
from io import StringIO
import matplotlib.pyplot as plt
import matplotlib.gridspec as gridspec
from matplotlib.widgets import Slider, CheckButtons

text = '''Product,Count
Pen,10
Pencil,15
Book,10'''
df = pd.read_csv(StringIO(text))

fig, ax = plt.subplots()
gs = gridspec.GridSpec(
    nrows = 2,
    ncols = 2,
    figure = fig,
    wspace = 0.3,
    hspace = 0.6,
    height_ratios = [2,1]
    )
ax.set_position(gs[0,:].get_position(fig))

axMinCount = fig.add_subplot(gs[1,0])
axProducts = fig.add_subplot(gs[1,1])

labels = ('Pen', 'Pencil', 'Book')
minimum = 5
actives = [True, True, True]

df.loc[actives & (df['Count'] >= minimum)].plot(
    x = 'Product', y = 'Count', kind = 'bar', ax = ax, legend = False
    )

sMinCount = Slider(axMinCount, 'Min Count', 0, 20, valinit = minimum, valstep = 1)
cProducts = CheckButtons(axProducts, labels, actives)


def update(val):
    minimum = sMinCount.val
    df_filtered = df.loc[actives & (df['Count'] >= minimum)]
    if not df_filtered.empty:
        df_filtered.plot(
        x = 'Product', y = 'Count', kind = 'bar', ax = ax, legend = False
        )
    else:
        ax.cla()

def check(label):
    index = labels.index(label)
    actives[index] = not actives[index]
    df_filtered = df.loc[actives & (df['Count'] >= minimum)]
    if not df_filtered.empty:
        df_filtered.plot(
        x = 'Product', y = 'Count', kind = 'bar', ax = ax, legend = False
        )
    else:
        ax.cla()
    
sMinCount.on_changed(update)
cProducts.on_clicked(check)

plt.show()

使用各种过滤设置，结果如下所示：

ipywidgets（Jupyter 笔记本）

我建议也尝试ipywidgets，它的用户界面比matplotlib.widgets. 您可以阅读有关使用 Interact 的更多信息。使用您的最小示例：

import pandas as pd
from io import StringIO
from ipywidgets import interact

text = '''Product,Count
Pen,10
Pencil,15
Book,10'''
df = pd.read_csv(StringIO(text))

# This is a wrapper of the function that follows, providing the interactive input
@interact(MinCount = (0, 20, 1), pen = True, pencil = True, book = True)
# Note that in the core function below, you can set the starting values
def plotter_fun(MinCount = 0, pen = True, pencil = True, book = True):   
    # Filter the data using the interactive input
    df_filtered = df.loc[(pen, pencil, book) & (df['Count'] >= MinCount)]
    # If all data has been filtered out, announce it
    if df_filtered.empty:
        print('No data to show.')
    # Otherwise plot
    else:
        df_filtered.plot(x = 'Product', y = 'Count', kind = 'bar')

具有各种过滤设置的结果如下所示：

当然，配置布局等有很多选项。

此解决方案旨在主要在Jupyter Notebook 中工作，但如果您想将此功能嵌入其他地方，您可以阅读有关在 Notebook 之外的其他上下文中嵌入 Jupyter 小部件的信息。